Once again I would like to thank everybody for their input. I downloaded and tested the FLL that Igor built. It is faster (in a particularly effective way) than the way I was doing it. The blocks of text are about 5k to 20K in length when encoded. Even though the payload from server to browser is a bit larger when sending the fully "escaped" FLL output rather than the strconv() output, the browser is substantially faster at actually rendering the web page. Because frequently the page rendering is actually a refresh of the same page, the benefit of the FLL is elimination of what I call a "flicker" when the page is re-rendered. It seems that the browser first clears its client area, then proceeds with parsing/rendering. This was causing a fleeting pause (with a blank client area) for a fraction of a second while the browser's javascript interpreter was iterating through the strconv() output and recomposing fully-escaped characters. (Much more noticeable on my 800 Mhz P3 notebook than on my 2.4 Ghz desktop). Using the FLL, the browser is receiving fully-escaped (hexed) text and passing this text directly to its built-in unescape routine, and rendering is much faster. Very snappy. Many thanks!